Initializing Chat usage: demo.py [-h] --cfg-path CFG_PATH [--options OPTIONS [OPTIONS ...]] demo.py: error: unrecognized arguments: --gpu-id 0

[srtj]

Initializing Chat usage: demo.py [-h] --cfg-path CFG_PATH [--options OPTIONS [OPTIONS ...]] demo.py: error: unrecognized arguments: --gpu-id 0

不知道为什么报这个错误

[TsuTikgiau]

Hello! Do you use the latest code? We updated the code yesterday to include the argument gpu-id to specify the GPU device

[srtj]

是的， 我用的旧版本，现在已经更新了，但是又报了一个“RuntimeError: PytorchStreamReader failed reading zip archive: failed finding central directory”不知道怎么回事

[TsuTikgiau]

Can you share the whole error message?

[srtj]

Can you share the whole error message?

E:\github\MiniGPT-4\demo.py:60 in │ │ │ │ 57 model_config = cfg.model_cfg │ │ 58 model_config.device_8bit = args.gpu_id │ │ 59 model_cls = registry.get_model_class(model_config.arch) │ │ ❱ 60 model = model_cls.from_config(model_config).to('cuda:{}'.format(args.gpu_id)) │ │ 61 │ │ 62 vis_processor_cfg = cfg.datasets_cfg.cc_sbu_align.vis_processor.train │ │ 63 vis_processor = registry.get_processor_class(vis_processor_cfg.name).from_config(vis_pro │ │ │ │ E:\github\MiniGPT-4\minigpt4\models\mini_gpt4.py:243 in from_config │ │ │ │ 240 │ │ max_txt_len = cfg.get("max_txt_len", 32) │ │ 241 │ │ end_sym = cfg.get("end_sym", '\n') │ │ 242 │ │ │ │ ❱ 243 │ │ model = cls( │ │ 244 │ │ │ vit_model=vit_model, │ │ 245 │ │ │ q_former_model=q_former_model, │ │ 246 │ │ │ img_size=img_size, │ │ │ │ E:\github\MiniGPT-4\minigpt4\models\mini_gpt4.py:74 in init │ │ │ │ 71 │ │ for layer in self.Qformer.bert.encoder.layer: │ │ 72 │ │ │ layer.output = None │ │ 73 │ │ │ layer.intermediate = None │ │ ❱ 74 │ │ self.load_from_pretrained(url_or_filename=q_former_model) │ │ 75 │ │ │ │ 76 │ │ if freeze_qformer: │ │ 77 │ │ │ for name, param in self.Qformer.named_parameters(): │ │ │ │ E:\github\MiniGPT-4\minigpt4\models\blip2.py:77 in load_from_pretrained │ │ │ │ 74 │ │ │ cached_file = download_cached_file( │ │ 75 │ │ │ │ url_or_filename, check_hash=False, progress=True │ │ 76 │ │ │ ) │ │ ❱ 77 │ │ │ checkpoint = torch.load(cached_file, map_location="cpu") │ │ 78 │ │ elif os.path.isfile(url_or_filename): │ │ 79 │ │ │ checkpoint = torch.load(url_or_filename, map_location="cpu") │ │ 80 │ │ else: │ │ │ │ C:\Users\zj136\anaconda3\envs\minigpt4\lib\site-packages\torch\serialization.py:705 in load │ │ │ │ 702 │ │ │ # If we want to actually tail call to torch.jit.load, we need to │ │ 703 │ │ │ # reset back to the original position. │ │ 704 │ │ │ orig_position = opened_file.tell() │ │ ❱ 705 │ │ │ with _open_zipfile_reader(opened_file) as opened_zipfile: │ │ 706 │ │ │ │ if _is_torchscript_zip(opened_zipfile): │ │ 707 │ │ │ │ │ warnings.warn("'torch.load' received a zip file that looks like a To │ │ 708 │ │ │ │ │ │ │ │ " dispatching to 'torch.jit.load' (call 'torch.jit.loa │ │ │ │ C:\Users\zj136\anaconda3\envs\minigpt4\lib\site-packages\torch\serialization.py:242 in init │ │ │ │ 239 │ │ 240 class _open_zipfile_reader(_opener): │ │ 241 │ def init(self, name_or_buffer) -> None: │ │ ❱ 242 │ │ super(_open_zipfile_reader, self).init(torch._C.PyTorchFileReader(name_or_bu │ │ 243 │ │ 244 │ │ 245 class _open_zipfile_writer_file(_opener):

[TsuTikgiau]

it looks like there are some issue with the cached weights of Q-Former. Can you maybe find the cached weight, deleted it and try again?

[srtj]

Q-Former

it looks like there are some issue with the cached weights of Q-Former. Can you maybe find the cached weight, deleted it and try again?

Q-former 是需要下载的吗

[zhaodice]

删除 ~/.cache/torch/hub/checkpoints/eva_vit_g.pth 就行了